Skip to content

orbits64/project-synapse

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 

Repository files navigation

Project Synapse

Orbits Systems — 2026

A controlled experiment testing whether a dedicated reasoning component outperforms a standard neural network baseline on multi-step logic problems — specifically problems requiring generalization to entities never seen during training.


The Question

Does architectural specialization matter, or is a bigger standard model always the answer?

We isolated a ternary-weight reasoning module and compared it against an MLP baseline of comparable size. Both were trained on top of a frozen GPT-2 backbone, so the only variable was the classifier architecture.


Results

Run Baseline (MLP) Reasoning Component Gap
1 75.1% 87.3% +12.2%
2 68.5% 88.1% +19.6%

Test set used entirely different entity names than training. The reasoning component generalized. The baseline overfit and stalled.


How It Works

  • GPT-2 (frozen) converts puzzle text to numerical representations. Neither model trains GPT-2 — it is a shared, fixed input layer.
  • Baseline is a 3-layer MLP (214K parameters) — fully connected layers, no memory, no iteration.
  • Reasoning Component (272K parameters) runs iterative steps over explicit memory cells using ternary weights {-1, 0, +1}, with a confidence gate that exits early on easier problems and uses all steps on harder ones.

Run It

pip install torch transformers
python synapse.py

Reproduces the full experiment. Results print at the end. Takes about 2-3 minutes on GPU, longer on CPU.


Paper

Not All Weights Are Equal: A heterogeneous design for a reasoning-first model Beniven, Orbits Systems, 2026

Available on arXiv.


License

MIT — use it, build on it, cite it if it helps you.

© 2026 Orbits Systems

About

Controlled experiment comparing a ternary-weight reasoning module against an MLP baseline on multi-step logic. Orbits Systems research.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages